Re: [PATCH 2/3] staging: greybus: use inline function for macros

Uwe Kleine-König <u.kleine-koenig@xxxxxxxxxxxxxx> · Tue, 21 Mar 2023 17:26:29 +0100

On Tue, Mar 21, 2023 at 04:59:49PM +0100, Julia Lawall wrote:
> 
> 
> On Tue, 21 Mar 2023, Uwe Kleine-König wrote:
> 
> > Hello,
> >
> > just some nitpicks:
> >
> > On Tue, Mar 21, 2023 at 01:04:33AM +0200, Menna Mahmoud wrote:
> > > Convert `to_gbphy_dev` and `to_gbphy_driver` macros into a
> > > static inline function.
> > >
> > > it is not great to have macro that use `container_of` macro,
> >
> > s/it/It/; s/macro/macros/; s/use/use the/;
> >
> > > because from looking at the definition one cannot tell what type
> > > it applies to.
> > > [...]
> > > -#define to_gbphy_dev(d) container_of(d, struct gbphy_device, dev)
> > > +static inline struct gbphy_device *to_gbphy_dev(const struct device *d)
> >
> > drivers/staging/greybus/gbphy.c always passes a variable named
> > "dev" to this macro. So I'd call the parameter "dev", too, instead of
> > "d". This is also a more typical name for variables of that type.
> 
> I argued against that.  Because then there are two uses of dev
> in the argument of container_of, and they refer to completely different
> things.  It's true that by the way container_of works, it's fine, but it
> may be misleading.

Hmm, that seems to be subjective, but I have less problems with that
than with using "d" for a struct device (or a struct device_driver).
I'd even go so far as to consider it nice if they are identical.

Maybe that's because having the first and third argument identical is
quite common:

	$ git grep -P 'container_of\((?<ident>[A-Za-z_0-9-]*)\s*,[^,]*,\s*\g{ident}\s*\)' | wc -l
	5940

which is >44% of all the usages

	$ git grep -P 'container_of\((?<ident>[A-Za-z_0-9-]*)\s*,[^,]*,\s*(?&ident)\s*\)' | wc -l
	13362

Best regards
Uwe

-- 
Pengutronix e.K.                           | Uwe Kleine-König            |
Industrial Linux Solutions                 | https://www.pengutronix.de/ |